Forecasting Paid Social

13 August, 2020


Josh Duggan

Josh Duggan

Working in paid media for over 8 years. Specialises in Google Shopping and paid social. Spoken at various eCommerce events around the world about paid media and Google Shopping.

For a paid social campaign to meet the expectations of all stakeholders, it's essential to spend time on forecasting. This guide explains how to approach this crucial aspect of your campaign.

When planning paid social campaigns, forecasting will be one of the most important initial considerations. The main questions will be:

  • How much do I need to spend?

  • How many more sales/ revenue will these campaigns generate?

First, there a few things to consider:

  • Too low a budget will hinder your ability to generate learnings and results

  • The correct objectives need to be considered for each paid social strategy i.e. the objective of prospecting campaigns (Upper funnel) may be clicks to site/ add to cart whereas remarketing (High intent, Lower funnel) would be purchases.

  • Measurement - how will you report back on and measure performance.

  • Start with either a budget, so you can work backwards and calculate how many sales will be generated and therefore revenue, or alternatively, start with a sales/revenue goal and calculate how much spend you will need to reach an audience large enough to meet those objectives.

  • Most importantly though is the consideration between new and returning customers. You will want to consider whether the campaign is being served purely for new customer acquisition, or whether you want to target old and existing customers also. If the primary goal is sales - you will likely maximise your spend on returning customers (more efficient) and then scale up to broader audiences depending on budget.

Gather data on key performance metrics

The basis of your forecast will rely on some key assumptions: audience size (impressions), CTR (click-through rate), CPM (Cost per thousand impressions), CVR (Conversion rate), AOV (Average order value) and CPC (Cost per click).

If you have had similar campaigns running before, this is where you should start. However, make sure you are using data from a campaign that was using similar audience strategies to your new ones. For instance, your historical Paid social campaign which was targeting your remarketing audiences will have a higher CPM and conversion rate compared to your new prospecting campaign targeting users who may be interested in luxury fashion. Typically, impression and click costs are more expensive for smaller audience pools (in this instance remarketing).

If you do not have historical data, you can find benchmarks from similar clients or from online sources i.e. Ad Stage publishes engagement metrics regularly for Facebook, Twitter, LinkedIn, and YouTube. You can then overlay this with conversion rate data which you can find in Google Analytics. You can use your overall site average conversion rate but make sure you weight averages based on your campaign strategy i.e. prospecting vs remarketing.

Understand your audience sizes & estimate frequency

For remarketing audiences, you want to maximise your reach and, in most cases, prioritise this audience over prospecting (unless your objective is to drive new users to site, of course). It is important to understand your current audience size, and how often you want to show your ads. Over time, you will gather data which will enable you to understand when the frequency is too high for your campaign, i.e. when CTR starts falling and CPA starts creeping up. We have a proprietary Facebook app which will adjust remarketing budgets as the frequency gets too high or if the CPA starts to drop for example. As the budget is lowered, Facebook will reduce the spend and how often users see ads. If you are using CBO also, Facebook is pretty good at reducing spend first to the customers less likely to buy (improving ROI).

Over a 2 week campaign, you may want to start off with 2-4 as a foundation for your forecast. 2 for prospecting and 4 for remarketing, for instance. This will vary massively depending on your industry and business. Retailers selling higher priced items will often see a longer purchase cycle and may need a higher frequency to drive the sale.

You will also want to factor in how your prospecting audiences will fuel the growth of your remarketing audience. As you prospect for new customers and drive them to the site, they will be pushed into your remarketing pool, increasing the size and spend there.

Start to build your forecast

So, now it is time to put this into action. Let us assume your objective is to drive growth in sales and revenue across Facebook over the course of August with a budget of £20K.

Start by mapping out the campaign strategies you are going to use, the specific audiences you are going to target, the campaign objective, audience size and frequency. You can use a similar template to the one below.

Campaign Strategy: Prospecting (Upper Funnel)

Audience: Interests: Luxury Fashion

Objective: Landing page views/ add to cart

Audience Size : X

Frequency: 2

Campaign Strategy: Look-a-Likes (Middle Funnel)

Audience: Look-a-likes: High Value Frequent purchasers

Objective: Add to cart/ Purchase

Audience Size : X

Frequency: 2

Campaign Strategy: Remarketing (Lower Funnel)

Audience: Remarketing: Users who have visited site/ added to cart last 14 days

Objective: Purchase

Audience Size : X

Frequency: 4 For each campaign strategy, you are then going to start adding in the key benchmarks you established earlier.

CPC: Benchmark Data

AOV: Benchmark Data

Conversion Rate: Benchmark Data

CTR: Benchmark Data

Now you can use your audience size and benchmarks to calculate the following metrics and complete your forecast, for each campaign strategy.

Clicks: Impressions * CTR

Impressions: Audience Size * Frequency

Spend: Clicks * CPC

Sales: Clicks * CVR

Revenue: Sales * AOV

ROI: Revenue / Spend

CPA: Spend / Sales Clicks, Impressions and Spend can be summed across all three campaign strategies.